Technology Portfolio Assessments Using a Multi-Objective Genetic Algorithm

This paper discusses the use of a Multi-Objective Genetic Algorithm to optimize a technology portfolio for a commercial transport. When incorporating technologies into a conceptual design, there are often multiple competing objectives that determine the benefits and costs of a certain portfolio. The set of designs that achieves the best values of these objectives will fall along a Pareto front that outlines the tradeoffs which will give the optimal design. Multi-Objective Genetic Algorithms determine the Pareto set by giving higher priority to dominant portfolios in the evolutionary optimization techniques of selection and reproduction. When determining the final Pareto optimal set it is important to ensure that only compatible portfolios of technologies are present. The technology compatibility represents a constraint on the optimization that can be handled internal to the genetic algorithms or through imposed logic that eliminates incompatibilities through a procedure known as gene correction. In order to properly benchmark the Multi-Objective Genetic Algorithm approach a comparison will be made with a traditional genetic algorithm utilizing a single fitness value incorporating the various objectives.